Merge branch 'master' into integrate-frontend

api/main.py

 from contextlib import asynccontextmanager
 
-from api.optimization import MultiProcessOptimizer
-from api.routers import runs
 from fastapi import BackgroundTasks, FastAPI, Request, Response
 from fastapi.staticfiles import StaticFiles
 from requests import request as make_request
 
+from api.optimization import MultiProcessOptimizer
+from api.routers import runs
+
 # see https://github.com/tiangolo/fastapi/issues/3091#issuecomment-821522932 and https://github.com/encode/starlette/issues/1094#issuecomment-730346075 for heavy-load computation
 
 DEBUG = True

task.py

@@ -5,15 +5,13 @@ from functools import lru_cache
 from statistics import mean
 from typing import Union
 
+from cli import argument_parser
 from datasets import Dataset, load_dataset
 from evaluate import load as load_metric
 from llama_cpp import LlamaGrammar, deque
-from torch.utils import data
-from tqdm import tqdm
-
-from cli import argument_parser
 from models import Llama2, LLMModel, OpenAI
 from opt_types import ModelUsage
+from tqdm import tqdm
 from utils import log_calls, logger
 
 SYSTEM_MESSAGE = """